Changelog
All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
0.8.1 (2021-08-30)
Added
- NEON implementation for aarch64 (#274)
0.8.0 (2021-08-29)
Added
- SSE2 autodetection support (#270)
Changed
- AVX2 performance improvements (#267, #267)
- MSRV 1.51+ (#267)
- Lock to
zeroize
<1.5 (#269)
Removed
xchacha
feature: all XChaCha*
types are now available by-default (#271)
0.7.3 (2021-08-27)
Changed
- Improve AVX2 performance (#261)
- Bump
cpufeatures
to v0.2 (#265)
0.7.2 (2021-07-20)
Changed
- Pin
zeroize
dependency to v1.3 (#256)
0.7.1 (2021-04-29)
Added
0.7.0 (2021-04-29) [YANKED]
Added
- AVX2 detection; MSRV 1.49+ (#200, #212)
XChaCha8
and XChaCha12
(#215)
Changed
- Full 64-bit counters (#217)
- Bump
cipher
crate dependency to v0.3 release (#226)
Fixed
rng
feature on big endian platforms (#202)- Stream-length overflow check (#216)
Removed
Clone
impls on RNGs (#220)
0.6.0 (2020-10-16)
Changed
- Rename
Cipher
to ChaCha
(#177) - Replace
block-cipher
/stream-cipher
with cipher
crate (#177)
0.5.0 (2020-08-25)
Changed
- Bump
stream-cipher
dependency to v0.7 (#161, #164)
0.4.3 (2020-06-11)
Changed
0.4.2 (2020-06-11)
Added
- Documentation improvements (#149)
Key
, Nonce
, XNonce
, and LegacyNonce
type aliases (#147)
0.4.1 (2020-06-06)
Fixed
- Links in documentation (#142)
0.4.0 (2020-06-06)
Changed
- Upgrade to the
stream-cipher
v0.4 crate (#121, #138)
0.3.4 (2020-03-02)
Fixed
- Avoid accidental
alloc
and std
linking (#105)
0.3.3 (2020-01-18)
Changed
- Replace macros with
Rounds
trait + generics (#100)
Fixed
- Fix warnings when building with
rng
feature alone (#99)
0.3.2 (2020-01-17)
Added
CryptoRng
marker on all ChaCha*Rng
types (#91)
0.3.1 (2020-01-16)
Added
- Parallelize AVX2 backend (#87)
- Benchmark for
ChaCha20Rng
(#87)
Fixed
- Fix broken buffering logic (#86)
0.3.0 (2020-01-15) [YANKED]
NOTE: This release was yanked due to a showstopper bug in the newly added buffering logic which when seeking in the keystream could result in plaintexts being clobbered with the keystream instead of XOR'd correctly.
The bug was addressed in v0.3.1 (#86).
Added
- AVX2 accelerated implementation (#83)
- ChaCha8 and ChaCha20 reduced round variants (#84)
Changed
- Simplify portable implementation (#76)
- Make 2018 edition crate; MSRV 1.34+ (#77)
- Replace
salsa20-core
dependency with ctr
-derived buffering (#81)
Removed
byteorder
dependency (#80)
0.2.3 (2019-10-23)
Security
- Ensure block counter < MAX_BLOCKS (#68)
0.2.2 (2019-10-22)
Added
- SSE2 accelerated implementation (#61)
0.2.1 (2019-08-19)
Added
- Add
MAX_BLOCKS
and BLOCK_SIZE
constants (#47)
0.2.0 (2019-08-18)
Added
impl SyncStreamCipher
(#39)XChaCha20
(#36)- Support for 12-byte nonces ala RFC 8439 (#19)
Changed
- Refactor around a
ctr
-like type (#44) - Extract and encapsulate
Cipher
type (#43) - Switch tests to use
new_sync_test!
(#42) - Refactor into
ChaCha20
and ChaCha20Legacy
(#25)
Fixed
- Fix
zeroize
cargo feature (#21) - Fix broken Cargo feature attributes (#21)
0.1.0 (2019-06-24)